The useful thing about a background check, if you happen to be building a delivery startup in 2013, is that it tells you whether the driver can begin. The maddening thing is that it may take long enough to make the answer irrelevant. Daniel Yanisse and Jonathan Perichon were engineers at Deliv, watching the on-demand economy hurry into existence while an old industry shuffled paper behind it. They wanted an API. Nobody seemed to have made the one they needed. So they quit and built it.
There is a clean founder story here, and clean founder stories should make us suspicious. Two engineers spot a bottleneck. Y Combinator admits them to its Summer 2014 batch. Customers arrive from the booming gig economy. Checkr turns scattered courthouse information into software that can return answers quickly. Revenue follows. Yanisse later joked that early customers welcomed him so warmly he assumed selling was always like this. One imagines many founders reading that line with the tender irritation normally reserved for people who sleep well on airplanes.
The more interesting story begins when the software works. Checkr could produce a fast, accurate report and still watch a candidate lose an opportunity because an employer treated any record as a stop sign. The founders started speaking with applicants. They heard what an old offense, stripped of context, could do to a present-day chance at work. The product problem widened. Speed was useful. Accuracy was essential. Relevance and judgment were harder.
A robot builder chooses the faster planet
Yanisse grew up in Le Mans, France. His father came from Syria, his mother from Romania; they met at medical school in Romania under communism, left for France and repeated their medical education when their degrees were not recognized. Yanisse has called them his heroes, crediting their willingness to begin again for his own persistence. It is a family anecdote with more entrepreneurial substance than most airport-book aphorisms.
At EPFL in Lausanne, he studied microengineering and gravitated toward robots and computing. He worked part-time with Phonak Acoustic Implants. Between degrees, he landed an unpaid internship with the Intelligent Robotics Group at NASA Ames. He worked on Mars rover prototypes, discovered Silicon Valley and found roommates on Craigslist. The space program was thrilling, but slow. The Valley was fast. Yanisse returned to Switzerland determined to strengthen his software skills, shaping his master's work around computer science in partnership with Cisco before heading back to California.
He joined mobile startup Mogreet and then Deliv, where he reunited with Perichon. They tried several ideas. Most, Yanisse has cheerfully conceded, were probably bad. The one hiding inside their day job was plain enough: background screening was a complicated backend process, and complicated backend processes could be abstracted. Twilio had done it for communications. Stripe had done it for payments. AWS had done it for computing. Why not an API for background checks?
The question suited his taste. Yanisse has said he likes boring infrastructure. Boring, in this case, meant regulated, fragmented and consequential. Court records did not share a crisp schema merely because a developer wished they would. Names had to be matched. Offenses had to be classified. Compliance rules varied. Humans remained in the loop. The apparent dullness concealed a bramble patch, which is often where durable software companies go to feed.
When the report contains a person
The traditional industry sold protection through fear: a dangerous stranger might enter your business, so you must search the past. Yanisse wanted Checkr to hold another idea beside safety. A candidate with a record might also be qualified, rehabilitated and ready to work. An offense could be real yet unrelated to the job. Treating every record identically was administratively neat and humanly careless.
Checkr built tools for employers to filter records according to role-relevant criteria and for candidates to add context. It formed Checkr.org in 2021 to work on fair-chance employment. By 2024, the company said its hiring initiatives had unblocked more than 1.7 million candidates. That April, Yanisse received a Trailblazer Award at MIT's Writing the Code event for advocating fair-chance hiring.
Mission supplied a compass, not a waiver from tradeoffs. Yanisse has been candid about this. A business serving employers must protect customers and candidates, align social purpose with commercial value and stay humble about problems that also belong to law and government. The company revised its language over time to put safety and fairness together. The tension did not disappear. It became the work.
The hangover after “up and to the right”
Checkr's early curve was absurdly steep. In later conversations, Yanisse described reaching roughly a $1 million revenue run rate within months. The company grew with Uber, Lyft and other marketplaces, then expanded into conventional employers. Capital arrived in matching quantities, including a $250 million Series E in 2021 at a reported $4.6 billion valuation.
Hypergrowth is delightful right up to the invoice. Yanisse has spoken about employees joining and leaving almost immediately, executive hires that failed, cultural strain and the pressure that follows inflated expectations. He says it took more than a decade to assemble a leadership team that truly gelled. His public reflections contain little of the founder as infallible chess master. He has said he feels he made every mistake in the book. The saving clause is that he enjoys sharing the mistakes and getting better.
His operating phrase is “fly high, fly low.” A CEO has to see the long horizon, then descend into the gears of a product, team or customer problem. Altitude is not prestige; it is an instrument. This sounds obvious until a calendar fills with meetings at precisely the wrong height.
He also changed how he judged talent. Prestige signals had biased him toward celebrated universities and large technology companies. Over time, he came to value people who had developed hunger in tougher businesses, spoke plainly about failure and could operate without the cushion of a famous brand. The engineer who once trusted elegant proxies grew wary of proxies in hiring, including his own.
AI needs a window in the black box
Checkr used machine learning early to classify messy records, match identities and standardize information. Yanisse is optimistic about AI's ability to reduce error and expose human bias. He is equally clear that a model can absorb the prejudice already present in a process. His advice is practical: know where bias may enter, be cautious when AI touches those points, and resist the comfort of a black box.
Inside Checkr, he has encouraged experimentation from the bottom up. In 2025, employees could expense tools such as Cursor, Replit and Lovable, with Yanisse arguing that people across engineering, marketing and human resources should discover the useful workflows themselves. It is a very engineer-founder policy: distribute the lab, observe what works, then build structure around evidence.
The company's scope is now broader than employment screening. A 2026 update described four lines: Workforce for hiring, Truework for income and employment verification, Trust for real-time risk intelligence, and Personal for user-controlled identity and credentials. Checkr reported more than 120,000 customers and over $800 million in 2025 gross revenue. Its proposed abstraction is no longer merely a background-check API. It is a data layer for decisions about trust.
That ambition lands at an awkward moment. Generative AI makes resumes, voices, faces and credentials easier to fake. Yanisse has warned about deepfake interviews, state-linked hiring fraud and accelerating identity scams. Checkr sees a business opportunity in the disorder, including possible work in government verification. The expansion also raises the stakes. A platform that helps decide who may work, borrow, rent or receive a service must be accurate, contestable and legible to the person being described.
Permission, carried home
Yanisse remains attached to the Valley for a simple reason: it let an engineer become a CEO by starting something, rather than by patiently ascending somebody else's hierarchy. In 2025 he told fellow EPFL alumni that he wanted to help European entrepreneurs take more risks, build meaningful companies and enter the American market. “Dream Big. Go Big,” he said, a slogan with the uncomplicated cheer of a luggage sticker.
The aspiration completes a loop. A French student went to NASA, discovered an ecosystem, sharpened his software skills and returned with a taste for speed. He built the missing API and, through growth, found the problem nested inside the problem. Now he wants to carry some of Silicon Valley's permission back across the Atlantic.
Checkr's story is still being written, and trust is an unruly subject. The most persuasive part of Yanisse's journey is not that software can settle it. His own evolution suggests something more durable: abstractions are useful precisely because they force their builders to notice what refuses to be abstracted. A record is data. A decision is policy. A candidate is a person. Good infrastructure must remember all three.